Thematic Analysis of Publication Activity Among Academic and Teaching Staff: A Case Study of the Financial University

In the context of the rapid growth of scientific publications, the increase in interdisciplinary research and increased competition in the academic environment, the tasks of analyzing and visualizing scientific activity are becoming especially actual. Modern digital tools allow not only to track publication dynamics, but also to identify key research areas, as well as stable groups of authors that form scientific communities. One of the effective approaches in this area is a combination of topic modeling methods and network analysis based on graph theory. Scientific organizations often face the problem of lack of operational information about the internal structure of research activities: which topics are most actively developing, what are the connections between authors and teams, who acts as the “cores” of scientific communities. This is especially true for large universities, where hundreds of researchers work, creating a significant number of scientific papers. In such situation, manual analysis becomes impossible, and automated text processing and graph analytics methods come to the rescue. This article is devoted to the analysis of the publication activity of authors of the Financial University. The purpose of the study is to identify the topics of scientific papers and identify scientific communities to understand the development of research activities of higher education institutions by example of the Financial University. The study presents an approach to forming a data set of scientific publications of authors of the Financial University. Visualization of publication dynamics and keyword analysis were carried out, allowing to identify common trends. The BERTopic model was used to solve the problem of text clustering and determining publication topics. Identification of scientific communities was implemented through the construction and analysis of a co-authorship graph, which allows to identify groups of researchers actively collaborating within certain scientific areas.

1. Devitsyn I.N., Savin I.V. A tool for the analysis of scientific communities based on topic modeling and graph theory. Uspekhi kibernetiki. 2020;1(4):13–21. (In Russ.). DOI: 10.51790/2712-9942-2020-1-4-2

2. Cohan A., Feldman S., Beltagy I., Downey D., Weld D.S. SPECTER: Document-level representation learning using citation-informed transformers. arXiv preprint. Cornell University. 2020;arXiv:2004.07180. URL: https://arxiv.org/ abs/2004.07180 DOI: 10.48550/arXiv.2004.07180

3. Wang L., Yang N., Huang X., Yang L., Majumder R., Wei F. Multilingual E 5 Text Embeddings: A Technical Report. arXiv preprint. Cornell University. 2024;arXiv:2402.05672. URL: https://arxiv.org/abs/2402.05672 DOI: 10.48550/arXiv.2402.05672

4. Li X., Li J. AnglE-optimized Text Embeddings. arXiv preprint. Cornell University. 2023;arXiv:2309.12871. URL: https://arxiv.org/abs/2309.12871 DOI: 10.48550/arXiv.2309.12871